vs
1:14
overall capability wins (0 ties across 15 scored). Llama 4 Scout leads overall.
- Reasoning1:7
- Coding0:3
- Arena0:1
- Pricing2:0
- Speed2:0
- Specs0:3
Verdict
Llama 4 Scout leads coding (widest gap: +21.2 on LiveCode); Llama 4 Scout ranks higher on LMArena (+110 Elo); Llama 4 Scout offers ~78.1x larger context (10M vs 128K).
- Llama 4 Scout leads coding (widest gap: +21.2 on LiveCode)
- Llama 4 Scout ranks higher on LMArena (+110 Elo)
- Llama 4 Scout offers ~78.1x larger context (10M vs 128K)
- Llama 4 Scout allows ~8.0x more max output (33K vs 4K)
- Llama 3.1 8B is ~1.6x cheaper on a blended token basis than Llama 4 Scout
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Llama 3.1 8B
- $0.22
- Llama 4 Scout
- $0.33
Llama 3.1 8B is about 1.5x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Meta
Meta
Family
Llama / Muse
Llama / Muse
License
Llama 3.1 Community
Llama 4 Community
Open weights
Yes
Yes
Release
Jul 23, 2024
Apr 5, 2025
Knowledge cutoff
2023-12
-
API / provider
Together / Fireworks (ref.)
Together / Fireworks (ref.)
Modalities
text → text
text, image → text
Specs
Context window
128K
10M
Max output
4K
33K
Parameters
8B
109B (17B act.)
Pricing
Input $/1M
$0.18
$0.18
Output $/1M
$0.18
$0.59
Blended $/1M (3∶1)
$0.18
$0.28
Speed
tok/s
160
110
TTFT (s)
0.15
0.3
Reasoning
MMLU-Pro
48.3
74.3
GPQA Diamond
25.9
58.7
Humanity's Last Exam
5.3
3.8
AIME 2025
4.3
14.0
MATH-500
51.9
84.4
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
8.5
15.2
AA-LCR v1.1
18.0
27.7
CritPt
—
—
MMMU-Pro
—
52.9
IFBench
28.6
39.5
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
9.1
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
11.6
32.8
Terminal-Bench 2.1
1.5
3.7
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
32.8
—
SciCode
13.2
21.3
CursorBench
—
—
SWE-Rebench
—
6.6
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,211
1,322
| Metric | Llama 3.1 8B | Llama 4 Scout | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Meta | Meta | — |
| Family | Llama / Muse | Llama / Muse | — |
| License | Llama 3.1 Community | Llama 4 Community | — |
| Open weights | Yes | Yes | — |
| Release | Jul 23, 2024 | Apr 5, 2025 | — |
| Knowledge cutoff | 2023-12 | - | — |
| API / provider | Together / Fireworks (ref.) | Together / Fireworks (ref.) | — |
| Modalities | text → text | text, image → text | — |
| Specs | |||
| Context window | 128K | 10M | B +9.9M |
| Max output | 4K | 33K | B +29K |
| Parameters | 8B | 109B (17B act.) | B +101 |
| Pricing | |||
| Input $/1M | $0.18 | $0.18 | tie |
| Output $/1M | $0.18 | $0.59 | A +$0.41 |
| Blended $/1M (3∶1) | $0.18 | $0.28 | A +$0.10 |
| Speed | |||
| tok/s | 160 | 110 | A +50 |
| TTFT (s) | 0.15 | 0.3 | A +0.15 |
| Reasoning | |||
| MMLU-Pro | 48.3 | 74.3 | B +26.0 pts |
| GPQA Diamond | 25.9 | 58.7 | B +32.8 pts |
| Humanity's Last Exam | 5.3 | 3.8 | A +1.5 pts |
| AIME 2025 | 4.3 | 14.0 | B +9.7 pts |
| MATH-500 | 51.9 | 84.4 | B +32.5 pts |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 8.5 | 15.2 | B +6.7 pts |
| AA-LCR v1.1 | 18.0 | 27.7 | B +9.7 pts |
| CritPt | — | — | — |
| MMMU-Pro | — | 52.9 | — |
| IFBench | 28.6 | 39.5 | B +10.9 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 9.1 | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 11.6 | 32.8 | B +21.2 pts |
| Terminal-Bench 2.1 | 1.5 | 3.7 | B +2.2 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | 32.8 | — | — |
| SciCode | 13.2 | 21.3 | B +8.1 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | 6.6 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,211 | 1,322 | B +110 Elo |
Llama 3.1 8B
- Context
- 128K / 4K out
- Parameters
- 8B
- Price
- $0.18 / $0.18
- Speed
- 160 tok/s · 0.15s TTFT
- Modalities
- text → text
- License
- Llama 3.1 Community
Compact Llama 3.1 instruct model for edge, RAG, and high-throughput classification at sub-$0.20/1M rates.
Llama 4 Scout
- Context
- 10M / 33K out
- Parameters
- 109B (17B act.)
- Price
- $0.18 / $0.59
- Speed
- 110 tok/s · 0.3s TTFT
- Modalities
- text, image → text
- License
- Llama 4 Community
Llama 4 variant with an extreme 10M context window for long-document retrieval workloads.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.